Recently, both weight loss and weight gain were associated with an increased risk of cardiovascular disease (CVD) and all-cause mortality among over 1.5 million Korean patients with type 2 diabetes mellitus, indicating a clear U-shaped association1. While the poor prognosis in patients with weight gain was in line with the current paradigm of obesity, the increased risk of CVD and death in patients with weight loss seemed counterintuitive. The authors interpreted this as surprising because of the well-known benefits of weight loss to cardiometabolic profiles and provided several methodological explanations such as loss of lean mass, body weight fluctuation, or unintentional weight loss1. However, the dynamics of lipophilic pollutants stored in adipose tissue may be involved in the U-shaped association between weight change and poor prognosis among patients with type 2 diabetes mellitus. A crucial fact largely unknown to researchers and clinicians is that human adipose tissue is widely contaminated with a tremendous number of environmental pollutants2. Typical examples are compounds classified as persistent organic pollutants (POPs). These pollutants have common features, such as strong lipophilicity, resistance to biodegradation, bio-magnification in the food chain, and very long half-lives of several years to decades. Examples of POPs include organochlorine pesticides, polychlorinated biphenyls (PCBs), dioxins, and polybrominated diphenyl ethers. In addition to POPs, less lipophilic chemicals with short half-lives have also been detected in adipose tissues. Therefore, in modern society, human adipose tissue can be considered an organ that stores exogenous chemicals that are not easily metabolized or excreted from the body. Since the possible harms of high-dose individual compounds belonging to POPs to wild animals and humans have been reported since the 1960s and 1970s, the production and use of POPs – especially chlorinated POPs – have been banned or strictly controlled worldwide. As a result, current human exposure to these chemicals has decreased to very low levels. However, humans are still living with continuous exposure to low-dose POPs because of the contamination of food chain. Human exposure to POPs begins in the uterus due to the easy transfer of compounds from mother to infant and continues after birth through breastfeeding and the consumption of POP-contaminated food. Once lipophilic pollutants such as POPs have entered the human body, they are primarily stored in adipose tissue and are slowly released into circulation during triglyceride lipolysis. From the viewpoint of the whole body system, the storage of these chemicals to adipose tissue can be seen as protective because it reduces the amount of POPs reaching other critical organs. Recently, chronic exposure to low-dose POPs has been linked to the risk of many chronic diseases including type 2 diabetes mellitus and cardiovascular diseases in the general population3, 4. Although high doses of individual compounds belonging to POPs are well-known mitochondrial toxins, a recent human and in vitro study demonstrated that even low-dose POPs can impair mitochondrial function through the impairment of oxidative phosphorylation5. Therefore, chronic exposure to low-dose POPs might be involved in the development of a wide range of mitochondrial dysfunction-related diseases. Almost all human studies have used the concentrations of POPs in circulation as exposure biomarkers. In fact, serum concentrations of POPs are not determined by the recent exposure amount from external sources, but by the amount of POPs released from adipose tissue into circulation6. Therefore, the dynamics of POPs stored in adipose tissue are directly related to the findings on POPs from recent epidemiological studies3, 4. Two situations in which increased lipolysis of adipose tissue leads to increased concentrations of POPs in circulation exist: (1) obesity with insulin resistance, and (2) weight loss2. These two situations seem contradictory from the viewpoint of adipose tissue. However, both result in a greater release of POPs from adipocytes into the circulation. Therefore, weight gain and weight loss, the two situations related to a poor prognosis that were reported by Park et al.1, can be seen as factors that increase serum concentrations of POPs (Figure 1). The poor prognoses among those who lose weight deserves further discussion. As unintentional weight loss itself is a marker of poor prognosis, distinguishing unintentional weight loss from intentional weight loss is important. However, any type of weight loss, whether intentional or unintentional, releases more POPs from adipose tissue into the circulation. In detail, the release of POPs during unintentional weight loss would be more problematic than that during intentional weight loss, because intentional weight loss is often accompanied by changes in lifestyle consisting of a healthy diet and exercise. In fact, a healthy diet and regular exercise have been suggested as a practical way to counteract the harm imposed by low-dose POPs in humans7. Nonetheless, a POP-related disadvantage even among persons with intentional weight loss exists. The Action for Health in Diabetes (Look AHEAD) study, a large randomized controlled trial that investigated the effects of intensive intentional weight loss in overweight or obese patients with type 2 diabetes mellitus, suggested this possibility. In this study8, overweight or obese patients with type 2 diabetes mellitus with intensive intentional weight loss demonstrated a significant improvement in most known CVD risk factors in the short term. However, they failed to reduce CVD events in the long term, compared with the control group. Recently, the dynamics of POPs observed during intentional weight loss have been suggested as a possible mechanism for the unexpected results obtained from the Look AHEAD study9. Better survival among overweight or obese type 2 diabetes mellitus patients has been repeatedly observed in epidemiological studies10. This phenomenon, known as ‘the obesity paradox’, has been reported in patients with a wide range of diseases, including type 2 diabetes mellitus. Compared with other diseases, the obesity paradox in patients with type 2 diabetes mellitus appears more perplexing because being overweight or obese is a strong risk factor for the development of type 2 diabetes mellitus, and weight reduction is recommended for better glycemic control among overweight or obese patients. At present, the obesity paradox observed in type 2 diabetes mellitus patients has also been interpreted as the result of methodological limitations in epidemiological studies10, similar to the explanation for the unexpected relationship between weight loss and poor prognosis observed in the study by Park et al.1 However, a biological plausibility for obesity paradox exists: the role of adipose tissue as a storage site of POPs can be more important in patients who already suffer from the disturbance of homeostasis, compared with healthy persons. Since POPs can act as mitochondrial toxins, the storage of POPs in adipose tissue with natural lipid droplets is relatively safer than the presence of POPs in other critical organs. In particular, having a large amount of ‘healthy’ adipose tissue may be beneficial for patients. Taken together, there is a missing piece of knowledge in the field of research on obesity and weight change. The amount of POPs released from adipose tissue into the circulation increases among persons who have obesity with insulin resistance or lose weight regardless of intention. Although intentional weight loss through lifestyle modification is widely recommended for overweight or obese persons, any approach that does not consider the role of POPs may not lead to long-term benefits, due to trade-off effects. The issue of obesity becomes complicated when the presence of POPs stored in adipose tissue is recognized. The current prevailing approach to weight management, which exclusively focuses on fat mass reduction, may not be optimal. The development of innovative weight management strategies that consider both fat mass and POP dynamics is required. This work was supported by the National Research Foundation (grant number 2019R1A2C1008958), funded by the Ministry of Science and ICT of the Republic of Korea. The authors declare no conflict of interest.